Evaluating with bsz 1 tgt_len 256 ext_len 0 mem_len 0 clamp_len -1 using a sliding window.
Evaluating with bsz 1 tgt_len 256 ext_len 0 mem_len 0 clamp_len -1 using a sliding window.
Evaluating with bsz 1 tgt_len 256 ext_len 0 mem_len 0 clamp_len -1 using a sliding window.
245568 positions evaluated.
Time : 3673.21s, 14.97ms/segment
====================================================================================================
| test loss  3.98 | test ppl    53.395 
====================================================================================================
